📄 中文摘要
提出了一种基于聚类引导的超连接Mamba模型（mHC-HSI），旨在提高高光谱图像（HSI）分类的性能。该模型采用了新颖的聚类引导Mamba模块，基于流形约束超连接（mHC）框架，能够显著提升空间-光谱特征的学习能力。通过将复杂且异质的高光谱图像分解为更小的聚类，模型有效地提取了空间和光谱信息。此外，mHC-HSI在多个数据集上的实验结果表明，其分类精度优于传统的深度学习模型，展示了在高光谱图像分类任务中的潜在应用价值。
聚类引导的超连接Mamba模型用于高光谱图像分类
📄 English Summary
mHC-HSI: Clustering-Guided Hyper-Connection Mamba for Hyperspectral Image Classification
A clustering-guided hyper-connection Mamba model (mHC-HSI) is proposed to enhance hyperspectral image (HSI) classification performance. This model incorporates a novel clustering-guided Mamba module based on the manifold-constrained hyper-connection (mHC) framework, significantly improving spatial-spectral feature learning. By decomposing complex and heterogeneous hyperspectral images into smaller clusters, the model effectively extracts both spatial and spectral information. Experimental results on multiple datasets demonstrate that mHC-HSI outperforms traditional deep learning models, showcasing its potential application value in hyperspectral image classification tasks.